Search

BACKGROUND: In the contemporary food industry, accurate and rapid differentiation of oolong tea varieties holds paramount importance for traceability and quality control. However, achieving this remains a formidable challenge. This study addresses this lacuna by employing machine learning algorithms – namely support vector machines (SVMs) and convolutional neural networks (CNNs) – alongside computer vision techniques for the automated classification of oolong tea leaves based on visual attributes. RESULTS: An array of 13 distinct characteristics, encompassing color and texture, were identified from five unique oolong tea varieties. To fortify the robustness of the predictive models, data augmentation and image cropping methods were employed. A comparative analysis of SVM‐ and CNN‐based models revealed that the ResNet50 model achieved a high Top‐1 accuracy rate exceeding 93%. This robust performance substantiates the efficacy of the implemented methodology for rapid and precise oolong tea classification. CONCLUSION: The study elucidates that the integration of computer vision with machine learning algorithms constitutes a promising, non‐invasive approach for the quick and accurate categorization of oolong tea varieties. The findings have significant ramifications for process monitoring, quality assurance, authenticity validation and adulteration detection within the tea industry. © 2023 Society of Chemical Industry. [ABSTRACT FROM AUTHOR]

[Display omitted] • A data-driven framework for zero-defect manufacture of composite parts is proposed. • Accurate and fast simulation of the autoclave processing of a composite part of industrial complexity is achieved. • A statistical surrogate model is trained from model data and used to automatically design the part's caul plate. • Successful defect elimination is achieved in the physical demonstrator made using this caul plate. Manufacture-induced defects are a critical issue in composites applications that result in high volumes of material waste and costly experimental trials to help mitigate this. Advances in process modelling techniques have enabled the prediction of defects. However, the high computational cost of these tools limits their usefulness in an industrial context as they often struggle with the ever-increasing size of industrial structures, and present significant challenges to undertaking large optimisation problems. In this work, a recently proposed homogenisation scheme is used to accurately simulate the autoclave processing of a composite part of industrial complexity in a fraction of the time of other state-of-the-art processes. Physical manufacture of a demonstrator part is used to evaluate the model in terms of accuracy. Based on the observed good agreement (discrepancy of less than 1.5% in the thicker sections and less than 7% in the thinner region of the part), a data-driven optimisation of the part's caul plate has been conducted that aims to mitigate defects whilst keeping cost and material use low. Further manufacturing trials validate the proposed framework by demonstrating successful defect elimination and a 25% improvement in the maximum local deviation from ideal design. [ABSTRACT FROM AUTHOR]

Background: Viral infections are prevalent in human cancers and they have great diagnostic and theranostic values in clinical practice. Recently, their potential of shaping the tumor immune microenvironment (TIME) has been related to the immunotherapy of human cancers. However, the landscape of viral expressions and immune status in human cancers remains incompletely understood.Methods: We developed a next-generation sequencing (NGS)-based pipeline to detect viral sequences from the whole transcriptome and used machine learning algorithms to classify different TIME subtypes.Results: We revealed a pan-cancer landscape of viral expressions in human cancers where 9 types of viruses were detected in 744 tumors of 25 cancer types. Viral infections showed different tissue tendencies and expression levels. Multi-omics analyses further revealed their distinct impacts on genomic, transcriptomic and immune responses. Epstein-Barr virus (EBV)-infected stomach adenocarcinoma (STAD) and Human Papillomavirus (HPV)-infected head and neck squamous cell carcinoma (HNSC) showed decreased genomic variations, significantly altered gene expressions, and effectively triggered anti-viral immune responses. We identified three TIME subtypes, in which the "Immune-Stimulation" subtype might be the promising candidate for immunotherapy. EBV-infected STAD and HPV-infected HNSC showed a higher frequency of the "Immune-Stimulation" subtype. Finally, we constructed the eVIIS pipeline to simultaneously evaluate viral infection and immune status in external datasets.Conclusions: Viral infections are prevalent in human cancers and have distinct influences on hosts. EBV and HPV infections combined with the TIME subtype could be promising biomarkers of immunotherapy in STAD and HNSC, respectively. The eVIIS pipeline could be a practical tool to facilitate clinical practice and relevant studies. [ABSTRACT FROM AUTHOR]

Most existing Web service recommendation models based on machine learning do not fully consider the high-order features interaction between users and services and with poor interpretability. In this paper, an Interpretable Web Service Recommendation model based on Disentangled Representation Learning (WSR-DRL) is proposed. First of all, to make full use of the service description information to improve the accuracy of Web service recommendation, the features representation of service name is obtained by using BERT model, and the local and global features representation of service description information is further obtained by combining 2-D CNN and Bi-LSTM. Then the disentangled convolution neural network is used to generate the high-order interaction features between users and services, and the neighborhood routing algorithm is used to mine the latent factors in these features. That improves the accuracy of Web service recommendation and make it interpretable. Finally, in order to verify the effectiveness of the model, several groups of experiments are carried out on real data sets. The experimental results show that compared with latest models such as DMF, DeepFM, DKN, GCMC, NDCG model and WSR-MGAT model, the WSR-DRL model proposed in this paper shows better performance on Precision@10, Recall@10, F1@10 and NDCG@10 evaluation metrics. [ABSTRACT FROM AUTHOR]

• Proposed a multi-sensor system for detection of indoor fire incidents at early stage. • Machine learning using support vector machine for effective classification. • Improved accuracy with reduced false alarms for fire incident classification and detection. • Cost-effective for embedded implementation and satisfying industrial needs. Accurate detection and early warning of fire hazard are crucial for reducing the associated damages. Due to the limitations of smoke-based detection mechanism, most commercial detectors fail to distinguish the smoke from dust and steam, leading to frequent false alarms and costly evacuation unnecessarily. To tackle this issue, we propose a fast and cost-effective indoor fire alarm system for real-time early fire detection under various scenarios, whilst significantly reducing the false alarms. Multimodal sensors are integrated to acquire the data of carbon monoxide, smoke, temperature and humidity, followed by effective data analysis and classification. For ease of embedded implementation, the support vector machine (SVM) is found to outperform the Random Forests (RF), K-means, and Artificial Neural Networks (ANN). On a public dataset and our own dataset, the proposed system performs promising, with the values of the precision, recall, and F1 of 99.8%, 99.6%, and 99.7%, respectively. [Display omitted] [ABSTRACT FROM AUTHOR]

Because of the impending energy crisis and the environmental Impact of fossil fuels, researchers are actively looking for alternatives, such as Helium-3 on the Moon. Although it remains challenging to explore energies on the Moon due to the long physical distance, the lunar features, such as craters and rilles, can be the hotspots for such energy sources, according to recent studies. Thus, identifying lunar features, such as craters and rilles, can facilitate the discovery of Helium-3 on the Moon, which is enriched in such hotspots. However, previously, no computational method was developed to recognize the lunar features automatically for facilitating space energy discovery. In our research, we aim at developing the first deep learning method to identify multiple lunar features simultaneously for potential energy source discovery. Based on the state-of-the-art deep learning model, High Resolution Net, our model can efficiently extract semantic information and high-resolution spatial information from the input images, which ensures the performance for recognizing the lunar features. With a novel framework, our method can recognize multiple lunar features, such as craters and rilles, at the same time. We also used transfer learning to handle the data deficiency issue. With comprehensive experiments on three datasets, we show the effectiveness of the proposed method. All the datasets and codes are available online. The source code and the data are available at https://github.com/frankchen121212/High-Resolution-MoonNet • The first deep learning method to identify lunar features for energy discovery. • One computational method to predict multiple features automatically and simultaneously. • Deployment of cutting-edge machine learning techniques: HRNet and transfer learning. [ABSTRACT FROM AUTHOR]